242 
Psyche 
[June 
preceding day, with activity peaks corresponding to peaks of soil 
warmth. It seemed that foraging rates of the species were mediated 
by soil surface temperatures. 
On 19 October I tested this hypothesis by using a Thermos Brand 
Space Blanket (a highly reflective material) to shade a nest of 
D. antarcticus on a hot, sunny day. Figures 3a and 3b show activity 
patterns at two neighboring nests of similar size. Workers of the 
unshaded nest illustrate the expected pattern, but workers from 
the shaded nest continued foraging throughout the day. Figures 4a 
and 4b show a similar test with T. antarcticum. Figure 4a illus- 
trates activity of a single colony on 1 November; Figure 4b is the 
same nest, shaded, on 2 November. The shading yielded a marked 
reduction in foraging rate. The late afternoon peak is of workers 
carrying larvae and pupae and emigrating to a new nest site under 
unshaded stones about 1 m from the shaded site. 
Discussion 
The two species studied have similar food and foraging site prefer- 
ences. The observed differences in temporal foraging pattern are 
hypothesized to contribute significantly to the species’ coexistence. 
Soil surface temperature seems to be the proximate factor by which 
the foragers regulate their activity. Endogenous activity rhythms, 
often keyed to environmental factors, are known for many species. 
This study illustrates how variation in environmental factors can 
yield alteration of the activity rhythms. 
During the course of a year at this temperate latitude (33° S) 
it seems probable that Dorymyrmex antarcticus would have more 
foraging hours available to it, especially during spring, fall, and 
winter days, than would Tapinoma antarcticum. Perhaps the aggres- 
sive dominance of Tapinoma is a necessary requisite for survival 
in competition with other species. The rapid recruitment of a large 
number of rather small workers in Tapinoma may illustrate com- 
ponents of a foraging strategy that must yield successful foraging 
returns during activity periods that are more limited than those of 
a coexisting competitor. 
Acknowledgments 
This research has been supported by NSF grant GB 31 195 to 
R. K. Colwell. This paper is a contribution from the Structure 
of Ecosystems Subprogram, International Biological Program. I 
thank R. R. Snelling, who taxonomically determined the species 
studied, and E. Reid, who prepared the illustrations. For review 
of the manuscript I am grateful to R. K. Colwell, B. Holldobler 
and E. O. Wilson. 
